KSP Mission Control — Voice-Controlled CAPCOM

Talk to Mission Control like you're really flying. This mod connects your KSP flights to a voice-controlled CAPCOM that gives live go-calls, verifies mission milestones against real telemetry, and talks back with NASA-style radio chatter throughout your mission.

⚠️ Three things are required for this to work

1. The companion desktop app (.exe, included in this download) — handles voice recognition, AI responses, and text-to-speech
2. The KSP mod (.dll, included in this download) — goes in your GameData folder, talks to kRPC
3. kRPC — a separate mod, download and install this yourself. It's the live connection between KSP and Mission Control.

All three need to be installed together, none of them work on their own.

🔄 The app updates itself

The companion app updates itself automatically — install it once and you'll always have the latest version. The KSP mod (.dll) only needs re-downloading if a release specifically notes a mod change, which is rare.

Requirements

• Windows — the companion app is currently Windows-only
• An OpenAI API key — used for voice/speech processing

Features

• Push-to-talk voice commands, speak your reports and orders naturally
• Live go-calls during launch, burns, and orbital insertion
• Real-time telemetry verification for every milestone, with alerts when something's off-script
• NASA-style radio chatter and callsigns
• Multiple mission types: Orbit, Flyby, Land, Land & Return, Rendezvous

See it in action: Watch the demo

Source & issues: GitHub

Privacy

KSP Mission Control runs on your own machine and doesn't collect your data. There's no analytics, no tracking, and nothing is sent to the developer.

What's sent to OpenAI: The app uses OpenAI's API (with your own API key) to work: - Your microphone audio is sent to OpenAI to transcribe your spoken reports into text. - The app's CAPCOM dialogue is sent to OpenAI to generate the spoken voice. Your voice and this text are processed by OpenAI under their API data policy (per OpenAI, API data isn't used to train their models by default). Audio is processed on the fly and the temporary files are deleted after use — no recordings or transcripts are kept.

Your API key is stored encrypted on your own machine (using your operating system's secure storage) and is only ever sent to OpenAI.

Other network activity: - On first launch, the app downloads a small audio component (ffmpeg) from a public GitHub repository. - The app periodically checks GitHub for updates (this is how auto-update works).

Stored on your machine: only your encrypted API key, the pilot/CAPCOM names you enter, and a few app settings. Nothing is stored remotely or shared with anyone.